Artificial intelligence (AI) promises to completely transform everyday life—it’s even been called the Fourth Industrial Revolution. The opportunities AI offers are tremendous, and every time I’ve discussed them with revenue agencies around the world, the same questions arise: Where should we focus and how can we maximize the returns on investment?
I see three primary use cases offering strong ROI. First, robotic process automation (RPA) can bring huge efficiency and quality gains in back and middle offices (think tax processing, compliance and support functions). Second, analytics models leveraging machine learning could support more complex operational decisions. And third, virtual digital assistants (VDAs) could transform customer and citizen support.
As they look to move forward, revenue agencies need to consider five key elements:
The technology is ready for business, but an agile approach to adoption is required, as vendors are improving products all the time. A product should offer rapid ROI, as it will make sense to swap technologies frequently.
Data quality is a fundamental building block for RPA and VDA to deliver on their expectations.
The technology offers a number of flexible architecture deployments to run proof of concepts—from desktop-based (allowing fast and low-cost implementation) to large-scale factory deployment in virtualised environments (for resilience, scalability and standardisation).
The tech isn’t the real challenge—it’s all about how it’s implemented in business-critical operations. Agencies need a mature RPA factory prioritising target processes for deployment. A revenue business process model will help accelerate this.
Security and trust
AI deployments come with security and privacy implications—revenue agencies need to think how the public can be reassured about these issues. Central government programmes will manage regulations for AI use, but revenue agencies will play a key role in defining the rules of the game. Transparency will be the key.
People and skills
Adopting AI means embarking on organizational change. Staff need to be empowered and reassured that AI is here to support—not replace—them. A "service design" approach will be essential. So will a workforce transition plan, as employees move from repetitive, low value-adding tasks to focus on core tax activities. The right mix of resources is a combination of AI, business, IT and process specialists. Revenue agencies must also ensure they’re not left behind in the coming war for AI talent and resources—relationships with universities and apprenticeship programmes are a great way forward.
Business case and roll-out
The business case for automation is clear: RPA is likely to allow for up to 30 percent automation, machine learning for an additional 20 percent. But achieving this means designing and deploying operating models in an agile and progressive way.
The 2017 Accenture Technology Vision revealed 85 percent of businesses will invest heavily in AI technologies over the next three years. And Gartner highlights how AI could help cut operating costs by nearly a third. There’s no doubt it’s going to be a serious lever for change across most industries. Revenue agencies are clearly getting primed for AI. They have the business cases and the operational readiness. The tech is all set—Accenture’s own experience of implementing everything from proofs of concept to industrialised deployments shows RPA and VDAs are ready. The EU is even making funding available (up to €500m overall) to fast-track AI in public administration. It’s clear leaders in revenue agencies can now ignite an AI led transformation.
I’d like to thank the European Centre for Government Transformation and the Lisbon Council for organizing a very insightful roundtable, where we discussed this topic with CIOs from major financial services businesses and senior figures from government. It’s a conversation I’d like to continue, and I’d welcome your views in response to this blog and about how you see AI playing out for revenue agencies.
For more related thinking and content click here
See this post on LinkedIn: AI: A no-brainer for revenue agencies